There's one other function of journals that few people seem to be bringing up or addressing: that of filtering the flood of research. Hundreds of papers per wee

There's one other function of journals that few people seem to be bringing up or addressing: that of filtering the flood of research. Hundreds of papers per week are published a week in "biology" across dozens of journals, but in most weeks, 0-1 of those will be interesting to me. By scanning the titles in Science, Nature, and Cell, I'm much likelier to find the really cool stuff than if I were just given a list of all papers with a given tag. I'm not saying that this recommendation is an intrinsically hard problem, just that it doesn't seem to be getting as much play as the other aspects of the open science revolution.